Pseudomonas aeruginosa is a major cause of morbidity and mortality in the ICU, particularly among patients with Ventilator-Associated Pneumonia. Many isolates are multi-drug resistant and some isolates are resistant to all currently extant anti-infective agents. There are currently no new antibiotics in clinical development (in man) with novel mechanisms of action against Pseudomonas. Given the cycle time associated with new drug development, it is likely that no antibiotics with new mechanisms of action for this pathogen will arise for 5 - 7 years. Thus, we must generate new knowledge about how best to suppress antibiotic resistance for this pathogen. This will help preserve our current drugs while we await new agents. Also, when agents with new mechanisms of action become available they can be developed in an optimal fashion for resistance suppression, both as monotherapy and in combination. In this application (Specific Aim #1), we hypothesize that we can identify optimal doses and schedules of administration for monotherapy for resistance suppression by studying this pathogen in our Hollow Fiber Infection Model (HFIM) and fitting a large mathematical model to the HFIM data to identify these doses and schedules. We further hypothesize that different resistance mechanisms will alter optimal doses and schedules. We propose to study isogenic mutants of the wild-type P. aeruginosa PAO-1 isolate, each containing a defined resistance mechanism. These findings will be bridged to man through use of Monte Carlo simulation (MCS). In Specific Aim #2, we propose to examine drugs in combination chemotherapy. We have developed a completely novel mathematical model that allows description of the impact of two drug combination chemotherapy on isolates of P. aeruginosa. This model is a mixture model and allows the fitting of the model to the concentration-time course of both agents as well as to fit the model to the disparate changes over time wrought by the combination of agents on the susceptible and less-susceptible populations of organisms present. Robust identification of the parameters of this system will allow calculation of optimal combination regimens for resistance suppression. Such regimens will be bridged to man through the use of MCS, as above. The HFIM is an in vitro system. In Specific Aim #3, we will also validate these optimal and non-optimal regimens in a neutropenic mouse pneumonia model, employing the same isolates studied in vitro in Aims #1 and #2. In examining this, we will use "humanized" dosing for the regimens, so that differences between mouse and human pharmacokinetics will not drive an improper inference. This will be done for both mono- and combination therapy. Prospective validation experiments will be designed and carried out. Results will be compared with the in vitro findings and also bridged to man. In so doing robust principles will be defined for drug regimens that will suppress amplification of resistant mutant populations. Pseudomonas aeruginosa is a pathogen of major importance in the Intensive Care Unit and is often resistant to many or even all of the drugs in our therapeutic armamentarium. As no agents with a unique mechanism of action active against Pseudomonas are expected for at least 7 years, it is imperative to learn how to use our currently available agents in a way that suppresses emergence of resistance and keeps these agents active for our patients. We plan to do this by 1) understanding the optimal way to dose Pseudomonas-active drugs in our hollow fiber infection model (HFIM) to suppress resistance and delineate the impact of different resistance mechanisms on this process 2) understand how to administer these drugs in combination in the HFIM to optimally suppress resistance emergence 3) validate the in vitro findings from the HFIM in a mouse model of Pseudomonas aeruginosa pneumonia.